Search

Data objects with mixed numeric and categorical attributes are commonly encountered in real world. The k-prototypes algorithm is one of the principal algorithms for clustering this type of data objects. In this paper, we propose an improved k-prototypes algorithm to cluster mixed data. In our method, we first introduce the concept of the distribution centroid for representing the prototype of categorical attributes in a cluster. Then we combine both mean with distribution centroid to represent the prototype of the cluster with mixed attributes, and thus propose a new measure to calculate the dissimilarity between data objects and prototypes of clusters. This measure takes into account the significance of different attributes towards the clustering process. Finally, we present our algorithm for clustering mixed data, and the performance of our method is demonstrated by a series of experiments on four real-world datasets in comparison with that of traditional clustering algorithms.

In many applications, data objects are described by both numeric and categorical features. The k-prototype algorithm is one of the most important algorithms for clustering this type of data. However, this method performs hard partition, which may lead to misclassification for the data objects in the boundaries of regions, and the dissimilarity measure only uses the user-given parameter for adjusting the significance of attribute. In this paper, first, we combine mean and fuzzy centroid to represent the prototype of a cluster, and employ a new measure based on co-occurrence of values to evaluate the dissimilarity between data objects and prototypes of clusters. This measure also takes into account the significance of different attributes towards the clustering process. Then we present our algorithm for clustering mixed data. Finally, the performance of the proposed method is demonstrated by a series of experiments on four real world datasets in comparison with that of traditional clustering algorithms.

As one of the fundamental features, color provides useful information and plays an important role for face recognition. Generally, the choice of a color space is different for different visual tasks. How can a color space be sought for the specific face recognition problem? To address this problem, we propose a sparse tensor discriminant color space (STDCS) model that represents a color image as a third-order tensor in this paper. The model cannot only keep the underlying spatial structure of color images but also enhance robustness and give intuitionistic or semantic interpretation. STDCS transforms the eigenvalue problem to a series of regression problems. Then one spare color space transformation matrix and two sparse discriminant projection matrices are obtained by applying lasso or elastic net on the regression problems. The experiments on three color face databases, AR, Georgia Tech, and Labeled Faces in the Wild face databases, show that both the performance and the robustness of the proposed method outperform those of the state-of-the-art TDCS model.

Canonical correlations analysis (CCA) is often used for feature extraction and dimensionality reduction. However, the image vectorization of CCA breaks the spatial structure of the original image, and the excessive dimensions of vectors often cause the curse of dimensionality problem. In this paper, we propose a novel feature extraction method based on CCA in multi-linear discriminant subspace by encoding each action sample as a high-order tensor. An optimization approach is presented to iteratively learn the discriminant subspace by unfolding the tensor along different tensor modes, which shows that most of the underlying data structure, including the spatio-temporal information, is retained and the curse of dimensionality problem is alleviated by the use of the proposed approach. At the same time, an incremental scheme is developed for multi-linear subspace online learning, which can improve the discriminative capability efficiently and effectively. In addition, the nearest neighbor classifier (NNC) is employed for action classification. Experiments on the Weizmann database show that the proposed method outperforms the state-of-the-art methods in terms of accuracy and time complexity, and it is robust against partial occlusion.

Locality preserving projections (LPP) is a widely used manifold reduced dimensionality technique. However, it suffers from two problems: (1) small sample size problem and (2) the performance is sensitive to the neighborhood size k. In order to address these problems, we propose an exponential locality preserving projections (ELPP) by introducing the matrix exponential in this paper. ELPP avoids the singular of the matrices and obtains more valuable information for LPP. The experiments are conducted on three public face databases, ORL, Yale and Georgia Tech. The results show that the performances of ELPP is better than those of LPP and the state-of-the-art LPP Improved1.

This paper aims to address one of the many problems existing in current facial recognition techniques using tensor (TensorFace Algorithm and its extensions). Current methods rasterize facial images as vectors, which result in a loss of spatial structure information of facial images. In this paper, we propose a method called Sp-Tensor to extend TensorFace by applying the sub-pattern technique. Advantages of the proposed method include: (1) a portion of spatial structure and local information of facial images is preserved; (2) dramatically reduce the computation complexity than other existing methods when building the model. The experimental results demonstrate that Sp-Tensor has better performance than the original TensorFace and Sp-PCA1, especially for facial images with un-modeled views and light conditions.

Discriminant information (DI) plays a critical role in face recognition. In this paper, we proposed a second-order discriminant tensor subspace analysis (DTSA) algorithm to extract discriminant features from the intrinsic manifold structure of the tensor data. DTSA combines the advantages of previous methods with DI, the tensor methods preserving the spatial structure information of the original image matrices, and the manifold methods preserving the local structure of the samples distribution. DTSA defines two similarity matrices, namely within-class similarity matrix and between-class similarity matrix. The within-class similarity matrix is determined by the distances of point pairs in the same class, while the between-class similarity matrix is determined by the distances between the means of each pair of classes. Using these two matrices, the proposed method preserves the local structure of the samples to fit the manifold structure of facial images in high dimensional space better than other methods. Moreover, compared to the 2D methods, the tensor based method employs two-sided transformations rather than single-sided one, and yields higher compression ratio. As a tensor method, DTSA uses an iterative procedure to calculate the optimal solution of two transformation matrices. In this paper, we analyzed DTSA's connections to 2D-DLPP and TSA, theoretically. The experiments on the ORL, Yale and YaleB facial databases show the effectiveness of the proposed method.

Inferring gene regulatory networks from large-scale expression data is an important topic in both cellular systems and computational biology. The inference of regulators might be the core factor for understanding actual regulatory conditions in gene regulatory networks, especially when strong regulators do work significantly. In this paper, we propose a novel approach based on combining neuro-fuzzy network models with biological knowledge to infer strong regulators and interrelated fuzzy rules. The hybrid neuro-fuzzy architecture can not only infer the fuzzy rules, which are suitable for describing the regulatory conditions in regulatory networks, but also explain the meaning of nodes and weight value in the neural network. It can get useful rules automatically without factitious judgments. At the same time, it does not add recursive layers to the model, and the model can also strengthen the relationships among genes and reduce calculation. We use the proposed approach to reconstruct a partial gene regulatory network of yeast. The results show that this approach can work effectively.

In the post-genomic biology era, the reconstruction of gene regulatory networks from microarray gene expression data is very important to understand the underlying biological system, and it has been a challenging task in bioinformatics. The Bayesian network model has been used in reconstructing the gene regulatory network for its advantages, but how to determine the network structure and parameters is still important to be explored. This paper proposes a two-stage structure learning algorithm which integrates immune evolution algorithm to build a Bayesian network. The new algorithm is evaluated with the use of both simulated and yeast cell cycle data. The experimental results indicate that the proposed algorithm can find many of the known real regulatory relationships from literature and predict the others unknown with high validity and accuracy.

Credit scoring is very important in business, especially in banks. We want to describe a person who is a good credit or a bad one by evaluating his/her credit. We systematically proposed three link analysis algorithms based on the preprocess of support vector machine, to estimate an applicant's credit so as to decide whether a bank should provide a loan to the applicant. The proposed algorithms have two major phases which are called input weighted adjustor and class by support vector machine-based models. In the first phase, we consider the link relation by link analysis and integrate the relation of applicants through their information into input vector of next phase. In the other phase, an algorithm is proposed based on general support vector machine model. A real world credit dataset is used to evaluate the performance of the proposed algorithms by 10-fold cross-validation method. It is shown that the genetic link analysis ranking methods have higher performance in terms of classification accuracy.

Detecting the boundaries of protein domains is an important and challenging task in both experimental and computational structural biology. In this paper, a promising method for detecting the domain structure of a protein from sequence information alone is presented. The method is based on analyzing multiple sequence alignments derived from a database search. Multiple measures are defined to quantify the domain information content of each position along the sequence. Then they are combined into a single predictor using support vector machine. What is more important, the domain detection is first taken as an imbalanced data learning problem. A novel undersampling method is proposed on distance-based maximal entropy in the feature space of Support Vector Machine (SVM). The overall precision is about 80%. Simulation results demonstrate that the method can help not only in predicting the complete 3D structure of a protein but also in the machine learning system on general imbalanced datasets.

In this paper, a novel quantum swarm evolutionary algorithm (QSE) is presented based on the quantum-inspired evolutionary algorithm (QEA). A new definition of Q-bit expression called quantum angle is proposed, and an improved particle swarm optimization (PSO) is employed to update the quantum angles automatically. The simulated results in solving 0-1 knapsack problem show that QSE is superior to traditional QEA. In addition, the comparison experiments show that QSE is better than many traditional heuristic algorithms, such as climb hill algorithm, simulation anneal algorithm and taboo search algorithm. Meanwhile, the experimental results of 14 cities traveling salesman problem (TSP) show that it is feasible and effective for small-scale TSPs, which indicates a promising novel approach for solving TSPs.

Micro-expressions are brief involuntary facial expressions that reveal genuine emotions and, thus, help detect lies. Because of their many promising applications, they have attracted the attention of researchers from various fields. Recent research reveals that two perceptual color spaces (CIELab and CIELuv) provide useful information for expression recognition. This paper is an extended version of our International Conference on Pattern Recognition paper, in which we propose a novel color space model, tensor independent color space (TICS), to help recognize micro-expressions. In this paper, we further show that CIELab and CIELuv are also helpful in recognizing micro-expressions, and we indicate why these three color spaces achieve better performance. A micro-expression color video clip is treated as a fourth-order tensor, i.e., a four-dimension array. The first two dimensions are the spatial information, the third is the temporal information, and the fourth is the color information. We transform the fourth dimension from RGB into TICS, in which the color components are as independent as possible. The combination of dynamic texture and independent color components achieves a higher accuracy than does that of RGB. In addition, we define a set of regions of interests (ROIs) based on the facial action coding system and calculated the dynamic texture histograms for each ROI. Experiments are conducted on two micro-expression databases, CASME and CASME 2, and the results show that the performances for TICS, CIELab, and CIELuv are better than those for RGB or gray.

A new method was described for using a recurrent neural network with bias units to predict contact maps in proteins. The main inputs to the neural network include residues pairwise, residue classification according to hydrophobicity, polar, acidic, basic and secondary structure information and residue separation between two residues. In our work, a dataset was used which was composed of 53 globulin proteins of known 3D structure. An average predictive accuracy of 0.29 was obtained. Our results demonstrate the viability of the approach for predicting contact maps.

A dynamic fuzzy clustering method is presented based on the genetic algorithm. By calculating the fuzzy dissimilarity between samples the essential associations among samples are modeled factually. The fuzzy dissimilarity between two samples is mapped into their Euclidean distance, that is, the high dimensional samples are mapped into the two-dimensional plane. The mapping is optimized globally by the genetic algorithm, which adjusts the coordinates of each sample, and thus the Euclidean distance, to approximate to the fuzzy dissimilarity between samples gradually. A key advantage of the proposed method is that the clustering is independent of the space distribution of input samples, which improves the flexibility and visualization. This method possesses characteristics of a faster convergence rate and more exact clustering than some typical clustering algorithms. Simulated experiments show the feasibility and availability of the proposed method.

One of important cues of deception detection is micro-expression. It has three characteristics: short duration, low intensity and usually local movements. These characteristics imply that micro-expression is sparse. In this paper, we use the sparse part of Robust PCA (RPCA) to extract the subtle motion information of micro-expression. The local texture features of the information are extracted by Local Spatiotemporal Directional Features (LSTD). In order to extract more effective local features, 16 Regions of Interest (ROIs) are assigned based on the Facial Action Coding System (FACS). The experimental results on two micro-expression databases show the proposed method gain better performance. Moreover, the proposed method may further be used to extract other subtle motion information (such as lip-reading, the human pulse, and micro-gesture etc.) from video.

A novel equivalent genetic algorithm based on extended strings (ESGA) is presented in this paper. The ESGA is equivalent to the canonical genetic algorithms (CGAs) in optimization problems, whereas the manner of the group by group implementation makes it different from the conventional genetic algorithms. In this paper the conditions and the rate of convergence of the ESGA are analyzed theoretically. It is shown that the ESGA can converge to the global optimal solution and the average value of the times of reaching the optimal point is finite. An analytical expression for the average rate of convergence of the ESGA is also provided.

Community detection in complex networks is a topic of considerable recent interest within the scientific community. For dealing with the problem that genetic algorithm are hardly applied to community detection, we propose a genetic algorithm with ensemble learning (GAEL) for detecting community structure in complex networks. GAEL replaces its traditional crossover operator with a multi-individual crossover operator based on ensemble learning. Therefore, GAEL can avoid the problems that are brought by traditional crossover operator which is only able to mix string blocks of different individuals, but not able to recombine clustering contexts of different individuals into new better ones. In addition, the local search strategy, which makes mutated node be placed into the community where most of its neighbors are, is used in mutation operator. At last, a Markov random walk based method is used to initialize population in this paper, and it can provide us a population of accurate and diverse clustering solutions. Those diverse and accurate individuals are suitable for ensemble learning based multi-individual crossover operator. The proposed GAEL is tested on both computer-generated and real-world networks, and compared with current representative algorithms for community detection in complex networks. Experimental results demonstrate that GAEL is highly effective at discovering community structure., 6 pages, 2 figures, 1 table 2009 Fourth International Conference on Computer Sciences and Convergence Information Technology(ICCIT '09)

Tensor subspace transformation, a commonly used subspace transformation technique, has gained more and more popularity over the past few years because many objects in the real world can be naturally represented as multidimensional arrays, i.e. tensors. For example, a RGB facial image can be represented as a three-dimensional array (or 3rd-order tensor). The first two dimensionalities (or modes) represent the facial spatial information and the third dimensionality (or mode) represents the color space information. Each mode of the tensor may express a different semantic meaning. Thus different transformation strategies should be applied to different modes of the tensor according to their semantic meanings to obtain the best performance. To the best of our knowledge, there are no existing tensor subspace transformation algorithm which implements different transformation strategies on different modes of a tensor accordingly. In this paper, we propose a fusion tensor subspace transformation framework, a novel idea where different transformation strategies are implemented on separate modes of a tensor. Under the framework, we propose the Fusion Tensor Color Space (FTCS) model for face recognition.

Propose a novel approach of rule extracting based on attribute significance and decision classification (REBSC). Condition attributes are only discretized with their own features in local discretization theory, and the eventual rule set normally can be achieved after attribute reduction. The REBSC approach given in this paper makes the most of each condition attribute significance and classification object, generates minimum decision rules without attribute reduction. Experiment one fully demonstrates the reasonability of the REBSC algorithm. Experiment two further proves its validity and testifies the significance of breakpoint division.

Face recognition plays a important role in computer vision. Recent researches show that high dimensional face images lie on or close to a low dimensional manifold. LPP is a widely used manifold reduced dimensionality technique. But it suffers two problem: (1) Small Sample Size problem; (2)the performance is sensitive to the neighborhood size k. In order to address the problems, this paper proposed a Matrix Exponential LPP. To void the singular matrix, the proposed algorithm introduced the matrix exponential to obtain more valuable information for LPP. The experiments were conducted on two face database, Yale and Georgia Tech. And the results proved the performances of the proposed algorithm was better than that of LPP.

Recent research efforts reveal that color may provide useful information for face recognition. For different visual tasks, the choice of a color space is generally different. How can a color space be sought for the specific face recognition problem? To address this problem, this paper represents a color image as a third-order tensor and presents the tensor discriminant color space (TDCS) model. The model can keep the underlying spatial structure of color images. With the definition of n-mode between-class scatter matrices and within-class scatter matrices, TDCS constructs an iterative procedure to obtain one color space transformation matrix and two discriminant projection matrices by maximizing the ratio of these two scatter matrices. The experiments are conducted on two color face databases, AR and Georgia Tech face databases, and the results show that both the performance and the efficiency of the proposed method are better than those of the state-of-the-art color image discriminant model, which involve one color space transformation matrix and one discriminant projection matrix, specifically in a complicated face database with various pose variations.

The discretization method of continuous attributes based on decision attributes which is discussed in document [3] can’t consider some special breakpoints very carefully. The modified algorithm on discretization given in this paper will improve the recognition accuracy and decrease the number of breakpoints. Experiment one gives the result about recognition of tea taste signal based document [3]’s algorithm. Experiment two gives the result about recognition of tea taste signal based on modified algorithm on discrtization method. By comparison with experiment one and experiment two testified the superiority of algorithm in this paper.

The subspace transformation plays an important role in the face recognition. LPP, which is so-called the Laplacianfaces, is a very popular manifold subspace transformation for face recognition, and it aims to preserve the local structure of the samples. Recently, many variants of LPP are proposed. LPP is a baseline in their experiments. LPP uses the adjacent graph to preserve the local structure of the samples. In the original version of LPP, the local structure is determined by the parameters t (the heat kernel) and k (k-nearest neighbors) and directly influences on the performance of LPP. To the best of our knowledge, there is no report on the relation between the performance and these two parameters. The objective of this paper is to reveal this relation on several famous face databases, i.e. ORL, Yale and YaleB.

We present a Condensation Nucleus Clustering (CNC) method based on our study of SVC algorithm. In CNC, data points are mapped by means of a Gaussian kernel to a high dimensional feature space, where we search for the minimal enclosing sphere. We consider the inner data images as the rough estimate of each cluster's core, and they can be easily clustered. We name the groups of cluster result as Condensation Nucleus. Then assign the remaining data points into each cluster by linear discriminant analysis. Furthermore, we improve the CNC method to GCNC (Gradational Condensation Nucleus Clustering). In GCNC, the remaining data are assigned to each cluster gradationally. With the Condensation Nucleus bigger and the remaining data less, the Condensation Nucleus grow up to the final cluster results. We compare our methods with other similar clustering algorithm to demonstrate the performance of the proposed method on several datasets.

Fuzzy C-Means is one of the clustering algorithms based on optimizing an objective function. The selection of the initial parameters of the number and the initial cluster centers play an important influence in the performance of the FCM. This paper proposes a new FCM clustering algorithm with two stages. The proposed algorithm not only resolves the problem of the initial choice of the cluster center effectively, but also decreases the time when clustering large volume of data. The computer simulation results show the effectivity and the superiority of the new algorithm.

A method of multi-view human action recognition based on multi-scale features via tensor analysis is proposed. A series of silhouettes are transformed to a Serials-Frame image, from which the multi-scale features are extracted to construct the eigenSpace of a tensor, which named Serials-Frame Tensor(SF-Tensor). The SF-Tensor subspace analysis is applied to separate the variable views and people information to recognize different actions. Experiment results obtained show that the proposed method attains a good recognition rate and improves the efficiency significantly.

Advantages and disadvantages of two common algorithms frequently used in the moving target detection: background subtraction method and frame difference method are analyzed and compared in this paper. Then based on the background subtraction method, a moving target detection algorithm is proposed. The background image used to process the next frame image is generated through superposition of the current frame image and the current background image with a certain probability. This algorithm makes the objects which stay long time to be a part of the background after a certain period of time, but not be detected as a part of foreground. The experimental results show that this algorithm can detect moving targets more effectively and precisely.

A novel chaotic immune algorithm (CImmune) is proposed to implement for parameter selection of Support Vector Regression (SVR). After adding chaotic local searching to the artificial immune procedure for parameter optimization of SVR, this method takes the advantages of both Chaos Optimization Algorithm (COA) and Artificial Immune Algorithm (AIA) to improve the effect of SVR efficiently. From experiments by cross validation on the simulated data and concrete compressive strength dataset from UCI, the results show that the proposed method has good capability of searching optimum and jumping out the local optimum easily in SVR model.

Particle swarm optimization(PSO) algorithm is applied in relevance feedback(RF) of image retrieval, which is combined with user feedback process effectively, The proposed method makes users' understanding to the retrieval goal become clear gradually, it could search image database effectively and comprehensively, and at the same time it could avoid the contradiction of efficiency and retrieval effect caused by multi-feedback. Experiments show that the proposed algorithm is validity.

ERA is the acronym of environment, reactive rules and agent. The model is a new effective multi-agent cooperation framework, which has been successfully applied in a wide range of fields. This paper presents the concept of belief into multi agent system. Each agent stands for a feasible routing, it moves according to its own belief matrix in the sensor networks environment. Agent has the ability to evaluate and adjust its belief matrix according to the historical routing. Agents will interact their believes to each other when they meet at the same sensor node. This paper employs a large number of data and other classical models to evaluate the algorithm, the experimental results show that this model prolongs the lifetime of wireless sensor network by reducing energy dissipation, which validates the algorithm's feasibility and availability.

An Immune Particle Swarm Optimization (IPSO) for parameters optimization of Support Vector Regression (SVR) is proposed in this article. After introduced clonal copy and mutation process of Immune Algorithm (IA), the particle of PSO is considered as antibodies. Therefore, evaluated the fitness of particles by the Cross Validation standard, the best individual mutated particle for each cloned group will be selected to compose the next generation to get better parameters ***C *** of SVR. It can construct high accuracy and generalization performance regression model rapidly by optimizing the combination of three SVR parameters at the same time. Under the datasets generated from sincx function with additive noise and forest fires dataset, experimental results show that the new method can determine the parameters of SVR quickly and the gotten models have superior learning accuracy and generalization performance.

Immune algorithm is a rising algorithm which simulates the organism immune system by computer. One of the immune algorithms named clonal selection algorithm is widely used due to its adaptability, implicit parallelism and diversity. A fitness function is constructed in this paper by using protein folding restrictions, such as amino acids' hydrophobicity rule, protein’s secondary structure folding rule, degree and amount of contacts in contact map, and other special rules. The experiment results of the prediction of 200 non-homological proteins with different lengths show that, this improved clonal selection algorithm has good adaptability and high efficiency.

In this paper, a novel support vector clustering algorithm based on k-means (SVC-KM) is presented not only improve the SVC running speed, but also overcome the weaknesses of k-means algorithm. Firstly, SVC algorithm was employed to identify some samples as outliers and some others as intra-cluster points, so that it removed noise and extracted samples near to the cluster core. Then the proposed method used Minimum Spanning Tree Pruning (MSTP) strategy on those intra-cluster points to initialize the number of clusters and clustering centroids which would be the parameters for k-means algorithm. Finally, it run k-means algorithm on subset without outliers to obtain the clustering result and assigned outliers into the nearest cluster. Applying the proposed algorithm to several test datasets, the experiments results compared with initial SVC algorithm and classical k-means strongly validate the efficiency and feasibility of SVC-KM.

In this paper, we apply the least-square support vector machine (LS-SVM) to operon prediction of Escherichia coli (E.coli), with different combinations of intergenic distance, gene expression data, and phylogenetic profile. Experimental results demonstrate that the WO pairs tend to have shorter intergenic distances, higher correlation coefficient and much stronger relation of co-envoled between phylogenetic profiles. Also, we dealt with the data sets extracted from WOs? and TUBs?, processed the intergenic distances with log-energy entropy, de-noised the Pearson correlation coefficients of two genes expression data with wavelet transform, and computed the Hamming distances of two phylogenetic profiles. Then we trained LS-SVM using part of the data sets and tested the trained classifier model using the rest data sets. It shows that different combinations of features could affect the prediction results. When the combination of intergenic distance, gene expression data and phylogenetic profile is taken as the input of LS-SVM in the linear kernel type, good results can be obtained, of which the accuracy, sensitivity and specificity are 92.34%, 93.54%, and 90.73%, respectively.

The performance of support vector machines (SVM) drops significantly while facing imbalanced datasets, though it has been extensively studied and has shown remarkable success in many applications. Some researchers have pointed out that it is difficult to avoid such decrease when trying to improve the efficient of SVM on imbalanced datasets by modifying the algorithm itself only. Therefore, as the pretreatment of data, sampling is a popular strategy to handle the class imbalance problem since it re-balances the dataset directly. In this paper, we proposed a novel sampling method based on genetic algorithms (GA) to rebalance the imbalanced training dataset for SVM. In order to evaluating the final classifiers more impartiality, AUC (area under ROC curve) is employed as the fitness function in GA. The experimental results show that the sampling strategy based on GA outperforms the random sampling method. And our method is prior to individual SVM for imbalanced protein domain boundary prediction. The accuracy of the prediction is about 70% with the AUC value 0.905.

The L-K means hierarchy clustering is proposed to decrease the long time of searching in a huge scale face database by reducing retrieving range. The database is divided into several sub sets after clustering. One-step split is used to divide the massive sub set into a few of small sets. Data redundant technology and predefined thresh are worked on a class's edge elements which will be re-signed into some categories for correctly searching in sets. The experiment results show that the proposed methods can effectively improve the searching speed while ensuring similar retrieval accuracy to search in the whole database.

Making full use of image information with its own to extract features is the crucial problem in content based image retrieval (CBIR). In this paper, quotient space(QS) granularity computing theory is imported into image retrieval field, granularity thinking in image retrieval is clarified, and a novel image retrieval method is proposed. Firstly, aiming at the different behaviors under different granularities, color and texture features are obtained respectively under different granularities, different quotient spaces are achieve; secondly, do the attribute combination to the obtained quotient spaces according to the quotient space granularity combination principle; and then realize image retrieval using the combined attribute function. Comparing with methods adopting single attribute feature the image retrieval method based on quotient space granularity combination utilizes the image information with its own in a more effective way. The experimental results demonstrate the feasibility and validity of the proposed method.

In this paper, an incremental learning based color iris recognition method is proposed. The inner and outer boundaries of the iris are localized employing Canny edge detection method and Hough transform algorithm in the red channel and the blue channel respectively. This method exploits some fusion strategies to combine the matching scores between the test image and the templates of each class. The strategies used in this paper are min, max, sum, product and weighted sum. The experiments were carried out on UBIRIS v1.0. The experiments show that, in some scenarios that acquisition conditions change leading to possible decrease of performance, because of its characteristic of learning in practice, the incremental learning method can handle the problem effectively.

Making full use of image information with its own to extract features is the crucial problem in content based image retrieval (CBIR). In this paper, we import quotient space granularity computing theory into image retrieval field, clarify the granularity thinking in image retrieval, and a novel image retrieval method is proposed. Firstly, aiming at the different behaviors under different granularities, obtain color features under different granularities, achieve different quotient spaces; secondly, do the attribute combination to the obtained quotient spaces according to the quotient space granularity combination principle; and then realize image retrieval using the combined attribute function. Comparing with methods adopting single attribute feature the image retrieval method based on quotient space granularity combination utilizes the image information with its own in a more effective way. The experimental results demonstrate the feasibility and validity of the proposed method.

In this paper, a kind of object representation model global structure constraint is presented, in which objects are described as constellations of points satisfied with their intrinsic specific global structure constraints. The spatial relations among all the patches of small color variations are extracted as shape model and the representative color information of patches are encoded and clustered as color model with color cluster information. Then, in the searching phase, mutual information is used as measurement with optimal algorithm to locate target objects in images by finding out the exactly matched position. In the experiment, we tested the approach on a collection of human face images and the results demonstrated the approach is simple, effective and efficient.

Biometrics-based personal identification is regarded as an effective method for automatically recognizing a person's identity with a high confidence. This paper presents a novel approach for personal identification using weighting relative distance of key point scheme on hand images. In contrast with the existing approaches, this system extracts multimodal features, including hand shape and palmprint to facilitate the task of coarse-to-fine dynamic identification. Five hand geometrical features are used to guide the selection of a small set of similar candidate samples at the coarse level matching stage. In the fine level matching stage, the weighting relative distance of key point approach is proposed to extract palmprint texture.

In this paper, we first studied some fusion strategies for color images based iris recognition. The strategies include Min, Max, Sum, Product and Weighted Sum. The fusion takes place at the matching scores level of red, green and blue channels. The experiments were carried out on UBIRIS v1.0. No fusion strategy can guarantee an improvement compared to the red channel, which yields the best single channel performance. We present a semi-supervised learning algorithm for iris recognition. Two methods are employed to estimate the distance between a test image and a subject. Every test image is appended to a class as a new template according to the classification result. The experiments show that, in some scenarios leading to acquisition condition change, the semi-supervised learning method can improve the performance.

We present a novel object localization approach based on the global structure constraint model (GSC) and optimal algorithm. In GSC, Objects are described as constellations of points satisfied with their specific global structure constraints. The spatial relations among all the patches having stable color information and their representative color information around patches are encoded. Then, the searching algorithm, i.e. genetic algorithm, is used to locate target objects in images by finding out the exactly matched position. In the experiment, we tested the approach on a collection of human face images and the results demonstrated the approach is simple, effective and efficient.

Aiming at the color distributing feature of image, a feature extraction and image retrieval algorithm using Discrete Wavelet Transform (DWT) and Singular Value Decomposition (SVD) is proposed, which is based on child-block partition and area division. Firstly, divide the image into child-blocks, DWT the color value of each block, and extract the low frequency coefficients as the color feature of this block, and then divide the image into areas, some matrixes are composed with the DWT coefficients in this area, which are SVDed to gain the retrieval feature vector in order to compute the difference between images. Experiments show that the proposed algorithm reduces dimension well, obtains good recall and precision, and has good retrieval effect.